Machine translation is one of the most dynamically improving areas of the translation industry. MT engines are continuously getting trained and becoming more accurate, thus making the translation processes (in selected areas and topics) smoother and more cost-effective.
We have already covered the most important types of machine translation engines as well as the most common use cases where MT can come in handy for translators and localization departments alike.
In this article, we’d like to concentrate on a sub-field of machine translation: post-editing machine translation, also known as PEMT. PEMT is a process, often used by translators and companies which consists of four steps (according to Nimdzi):
- Preparation of the source text
First, the source text needs to be prepared for post-editing. This is also when you should decide which MT engine to use. - Testing and QA
Once you have your machine translation engine candidates, you can take a small portion of your source text and test how the different engines perform. - Post-editing
This is what most people think of when PEMT comes into the picture: the step where a translator evaluates and edits the target text. - QA
The final evaluation of the target text by translators and reviewers.
Now that you have an idea about how post-editing machine translation works, here are four frequently asked questions when it comes to PEMT, and, of course, the answers.
What kind of texts are the best candidates for PEMT?
One size definitely does not fit all—and not all texts are suitable for post-editing machine translation. It all depends on what the type of the text is as well as where you want to use the translated documents.
Also, please note that there are use cases where a plain machine-translated text suffices. Some manuals or instructions, or internally used documents where the only goal is a common understanding (where it’s not crucial for the translation to be perfectly accurate) usually belong in this category.
Post-editing machine translation can be a great choice in case of the following document types:
- Blog posts or press releases
- Technical texts
- Informal documents
- News articles
- Manuals, instructions
However, you must be careful and/or refrain from this method if your source is:
- UX/UI copy, since these usually consist of small segments and are heavily dependent on the context.
- Texts that require extensive experience and knowledge in a specific field (such as life science, engineering, or law)
- Marketing/advertising copy, where wordplay and wits are an important factor
Lastly, there are texts which simply don’t lend themselves to machine translation and PEMT at all, such as literary texts or creative concepts.
What advantages does PEMT have?
Let’s see how the translation/localization workflow can benefit from PEMT.
Speed
As mentioned in oneword’s article, using MT can significantly reduce turnaround time in case of large bodies of text, especially if you don’t have reference documents or a translation memory beforehand.
Price
Because of the time aspect as well as the pricing difference between PEMT and regular human translation, post-editing machine translation is generally more cost-effective.
Quality
With the right MT engine and the appropriate amount of post-editing, you can achieve the same high quality as you would by translating the text from scratch.
What are the challenges of PEMT?
Besides the obvious advantages, post-editing machine translation also poses some challenges to be dealt with.
Accuracy
Context cannot always be predicted by a machine translation engine. This means that it also cannot take into account the style of the source text or the cultural references contained in the original copy.
Not all texts are suitable for PEMT
As mentioned in the first part of the article, you have to be very careful when selecting what texts you want translated with the PEMT method. If you choose a text which is context-dependent or requires expertise in a certain subject, you could end up with a target text which requires extensive editing.
The more editing a text needs, the more the price goes up, and as we established before, higher quality is not necessarily promised.
What skills/qualifications do you need if you want to work with PEMT?
Of course, you can be a translator and also choose to take on PEMT jobs. However, different skills are needed for translating a text from scratch and post-editing machine translation.
You will be the one to ensure that the end result conforms to the quality standards of your client and the project, as well as make sure that it is as close to the original as possible in terms of meaning, style, format, etc.
For an MT post-editing job, you need at least these three things:
- Knowledge of the context of your source text
- Knowledge of the source as well as the target language
- Excellent editing skills
When working with machine-translated text, you will have to make informed decisions about the target text. In some cases, light editing of the machine-translated text is perfectly sufficient; however, sometimes you will have to ditch the translation and go back to the original text and use human translation instead.
As an MT post-editor, you will have to keep a close eye on the source and the target, and correctly evaluate and differentiate between these two instances.
Finally, you also need familiarity with the software (TMS or CAT-tool) that you are using for the project. In memoQ, one of the most useful features for PEMT is the creation of editing time reports.
“memoQ can record how much time you spend editing a document: memoQ records the editing time for each segment, in each role. There is an editing time for when you work on the document as a Translator, another editing time when you work as a Reviewer 2, and so on.”
If you’re curious about how you can take advantage of machine translation when working with memoQ, check out our MT integrations page.
Do your best work with memoQ—start your free trial today!
machine translation MTPE post-editing
![Zsófia Lelner](https://blog.memoq.com/hubfs/MicrosoftTeams-image%20(8).png)
Zsófia Lelner
Linguist turned content marketer, telling the story of memoQ.